Acceptance testing of a CT scanner with a knowledge-based iterative reconstruction algorithm

Abstract

Implementation of efficient computed tomography (CT) detectors, automatic exposure control and iterative reconstruction algorithms allows to obtain high quality images while minimizing absorbed dose to the patients, in consonance with the ALARA criterion. In this work, we present the results of the acceptance testing of a Philips iCT Elite scanner with the IMR reconstruction algorithm. A Catphan 500 phantom was used for this purpose. RESULTS: The laser positioning system was misaligned, while the gantry and table were aligned. The high-contrast spatial resolution was 21 lp/cm. Regarding low contrast resolution, a 3 mm pin with 0.3% contrast was observed using IMR for CTDIvol = 10.4 mGy, while a 5mm pin was observed using filtered back projection (FBP). Average uniformity was ± 3 Hounsfield Units (HU) and noise was <0.5%. It was verified that when using IMR reconstruction, the noise remained constant when the CTDIvol was reduced. Nominal and measured slice thicknesses were compared, resulting in the following values (measured/nominal): 4.82/5, 2.52/2.5, 1.24/1, 1.09/0.8. At 120 kVp, the effective energy was 70 keV. CT numbers were measured for different materials, with results in accordance with expected values. The contrast scale value resulted in 2·10^-4 cm^-1HU^-1. CONCLUSIONS: Image quality and general system’s functionality were evaluated, with satisfactory results. The results were taken as a reference for routine quality controls. The IMR reconstruction algorithm resulted in significantly improved low-contrast resolution and lower noise at low doses when compared with FBP. The impact of IMR technology on clinical images remains to be assessed.